Tekton Pruner

TO BE UPDATED

Event based pruning of Tekton resources

Getting Started

Tekton Pruner is a controller that continuously monitors a predefined configuration in a ConfigMap and removes PipelineRuns and TaskRuns according to the specified settings.

By default, the Pruner controller watches the ConfigMap tekton-pruner-default-spec. Additionally, it includes two other controllers that monitor every PipelineRun and TaskRun event in the cluster, continuously reconciling to remove them based on the ConfigMap settings.

Current Features:

1. Ability to set a cluster-wide pruning configuration that applies to all namespaces in the cluster.

   Pruning Configuration Variations (Prioritized from Highest to Lowest):

   • Pruning PipelineRuns and TaskRuns based on TTL (time-to-live) in seconds.ttlSecondsAfterFinished
   • Defining the number of successful PipelineRuns/TaskRuns to retain.successfulHistoryLimit
   • Defining the number of failed PipelineRuns/TaskRuns to retain.failedHistoryLimit
   • Setting a fixed number of PipelineRuns/TaskRuns to retain, regardless of status.historyLimit

   sample configMap depiciting the above settings that apply to all namespaces in the cluster

    apiVersion: v1
    kind: ConfigMap
    metadata:
        name: tekton-pruner-default-spec
        namespace: tekton-pipelines
    data:
        global-config: |
            ttlSecondsAfterFinished: 600 # 5 minutes
            successfulHistoryLimit: 3 
            failedHistoryLimit: 3
            historyLimit: 5  

Work In Progress Features:

2. Ability to set pruning configuration specific to a Namespace that can overrides the cluster-wide settings

3. Ability to set pruning configuration specific to PipelineRuns/TaskRuns that can be identified either by

   • Parent Pipleline/Task Name
   • Matching Labels
   • Matching Annotations This configurations override the Namespace specific configuration

Installation TO BE UPDATED

Prerequisite: you'll need Tekton Pipelines installed on your cluster before you install Pruner.

To install the latest version of Pruner to your Kubernetes cluster, run:

kubectl apply --filename ??release.yaml

To install a specific version of Chains, run:

kubectl apply -f ??release.yaml

To verify that installation was successful, wait until all Pods have Status Running:

kubectl get po -n tekton-pipelines --watch

NAME                                          READY   STATUS      RESTARTS   AGE
tekton-pruner-controller-5756bc7cb9-lblnx      1/1     Running      0        10s

Setup WIP

To view and edit the default configuartions:

• Set up any additional configuration

Tutorials WIP

To get started with pruner, try out our getting started tutorial.

Want to contribute - TO BE UPDATED

We are so excited to have you!

• See CONTRIBUTING.md for an overview of our processes
• See DEVELOPMENT.md for how to get started
• See ROADMAP.md for the current roadmap Check out our good first issues and our help wanted issues to get started!
• See releases.md for our release cadence and processes